훈련된 랜덤 포레스트 모델을 제공하고 ReAct 상호 작용을 위해 Bee 프레임워크와 통합하는 MCP 서버를 구축하는 방법에 대한 전체적인 연습입니다.
이 저장소를 복제하세요 git clone https://github.com/nicknochnack/BuildMCPServer
MCP 서버를 실행하려면cd BuildMCPServeruv venvsource .venv/bin/activateuv add .uv add ".[dev]"uv run mcp dev server.py
에이전트를 실행하려면 별도의 터미널에서 다음을 실행하세요.source .venv/bin/activateuv run singleflowagent.py
git clone https://github.com/nicknochnack/CodeThat-FastMLcd CodeThat-FastMLpip install -r requirements.txtuvicorn mlapi:app --reload
자세한 제작 방법은 여기에서도 확인할 수 있습니다.
MCP 클라이언트 구축(SingleFlow 에이전트에서 사용)
ML 서버를 구축하는 원본 비디오
👨🏾💻 저자: Nick Renotte 📅 버전: 1.x 📜 라이선스: 이 프로젝트는 MIT 라이선스에 따라 라이선스가 부여되었습니다.
This tool predicts whether an employee will churn or not, pass through the input as a list of samples. Args: data: employee attributes which are used for inference. Example payload
[{
'YearsAtCompany':10,
'EmployeeSatisfaction':0.99,
'Position':'Non-Manager',
'Salary:5.0
}]
Returns:
str: 1=churn or 0 = no churn| Name | Required | Description | Default |
|---|---|---|---|
| data | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It mentions the tool 'predicts' and returns a string, but lacks critical behavioral details like accuracy, confidence scores, model limitations, rate limits, or error handling. This is insufficient for a prediction tool with zero annotation coverage.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is moderately concise but could be better structured. It front-loads the purpose but includes an example that might be verbose. Sentences like 'pass through the input as a list of samples' are somewhat redundant. Overall, it's adequate but not optimally efficient.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no annotations, no output schema, and low schema coverage, the description is incomplete. It covers basic purpose and a parameter example but misses behavioral traits, usage context, and detailed output explanation. For a prediction tool, this leaves significant gaps in understanding its operation and reliability.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 0%, so the description must compensate. It adds value by explaining 'data' as 'employee attributes used for inference' and provides an example payload with specific fields. However, it doesn't fully document all required attributes or their types beyond the example, leaving gaps in parameter understanding.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'predicts whether an employee will churn or not' with the verb 'predicts' and resource 'employee'. It specifies the input format ('list of samples') and output meaning ('1=churn or 0=no churn'). However, without sibling tools, it cannot demonstrate differentiation, so it doesn't reach the highest score.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus alternatives, prerequisites, or limitations. It only states what the tool does without context for its application, such as when predictions are needed or what data is required beyond the example.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
With only one tool, there is no possibility of ambiguity or overlap between tools, as there are no other tools to compare it to. The tool's purpose is clearly defined and distinct by default.
With only one tool, naming consistency is inherently perfect, as there are no other tool names to be inconsistent with. The tool name 'PredictChurn' follows a clear verb_noun pattern.
A single tool is too few for a server described as 'ML Model Integration', which suggests a broader scope covering multiple models or operations. This feels thin and incomplete for the apparent domain.
The server is severely incomplete for ML model integration, as it only offers a churn prediction tool. There are significant gaps, such as no tools for training models, listing available models, updating models, or handling other common ML tasks.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
HiveCompute MCP Server — decentralized inference router for AI agents
MCP server exposing the Backtest360 engine API as tools for AI agents.
MCP server unifying ERPs, CRMs, APIs and knowledge base for Claude, ChatGPT and Gemini.
NeuralBrain MCP Server - RAG, Vector Memory, LLM Routing, Agent Identity, x402 Payments
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/nicknochnack/BuildMCPServer'
If you have feedback or need assistance with the MCP directory API, please join our Discord server